Vincent Fletcher
Current Work:
A Complete Thermodynamic Description of Alloys to Extreme Pressure and Temperature
In October 2022 I joined the HetSys CDT and began work on a PhD in physics. The aim of my project is to develop a first-principles theoretical and computational scheme that provides a complete thermodynamic description of alloys under both ambient and extreme conditions. By 'complete' I mean a technique that is able to compute the Helmholtz free energy of the system, which can then be used to determine properties such as the equation of state and phase boundaries together with properties across boundaries such as latent heats.
Click the link to my project page for further detailsLink opens in a new window.
My project is in collaboration with our industry partner AWE - Nuclear Security TechnologiesLink opens in a new window
Supervising my project is Dr Albert P. BartókLink opens in a new window, and Dr Livia B. PártayLink opens in a new window.
Background:
Prior to starting my PhD project, in July 2022 I received a first class masters degree in chemistry (MChem) from the University of Warwick. While undertaking my degree, my studies focussed on organic, physical, computational, and quantum chemistry. For my final year project I worked in the Habershon research groupLink opens in a new window with Dr Gareth RichingsLink opens in a new window, applying recent advancements in machine learning to ab-initio quantum dynamics calculations, to study the fluorescence properties of HINA and HPALink opens in a new window.
Publications:
Optimal Autonomous MLIP Dataset BuildingLink opens in a new window
V.G.Fletcher, A.P.Bartók, L.B.Pártay
We propose a novel approach for constructing training databases for Machine Learning Interatomic Potential (MLIP) models, specifically designed to capture phase properties across a wide range of conditions. The framework is uniquely appealing due to its ease of automation, its suitability for iterative learning, and its independence from prior knowledge of stable phases, avoiding bias towards pre-existing structural data. The approach uses Nested Sampling (NS) to explore the configuration space and generate thermodynamically relevant configurations, forming the database which undergoes ab-initio Density Functional Theory (DFT) evaluation. We use the Atomic Cluster Expansion (ACE) architecture to fit a model on the resulting database. To demonstrate the efficiency of the framework, we apply it to magnesium, developing a model capable of accurately describing behaviour across pressure and temperature ranges of 0-600 GPa and 0-8000 K, respectively. We benchmark the model's performance by calculating phonon spectra and elastic constants, as well as the pressure-temperature phase diagram within this region. The results showcase the power of the framework to produce robust MLIPs while maintaining transferability and generality, for reduced computational cost.
Conferences:
--2025--
HetSys Summer Conference, University of Warwick: UK, July
-Contributed Talk
AWE - Nuclear Security Technologies Student Conference: UK, April
-Contributed Talk
CECAM - Fulfilling the Multiscale Promise in Materials, CECAM-HQ EPFL Campus: Switzerland, March
-Contributed Talk
Lennard-Jones Centre MLIP Workshop, University of Cambridge: UK, January
-Contributed Talk
--2024--
GAP/(M)ACE Developers & Users Meeting, Fritz Haber Institute Max Planck Society: Germany, September
-Contributed Talk
Computational Molecular Science (CMS), University of Warwick: UK, September
-Contributed Talk
HetSys Summer Conference, University of Warwick: UK, July
-Contributed Talk
Machine Learning for Quantum Mechanics (ML4QM), Imperial College London: UK, May
--2023--
GAP/(M)ACE Developers & Users Meeting, University of Warwick: UK, September
HetSys Summer Conference, University of Warwick: UK, July
AWE - Nuclear Security Technologies Student Conference, University of Oxford: UK, April
--2022--
CASTEP Training Workshop, University of Oxford: UK, August
Teaching:
As a graduate teaching assistant I have taught in the following modules:
PX159Link opens in a new window- Physics Programming:
A first year undergraduate physics module teaching Python programming in a physics context.
PX911Link opens in a new window- Multiscale Modelling Methods and Applications 1
A postgraduate physics module teaching molecular dynamics, machine learning, and electronic structure theory.
Key Words:
Statistical Mechanics, Machine Learning, Thermodynamics, Ab-Initio, Density Functional Theory (DFT), Phase Boundaries, Crystal Structures, Quantum Chemistry, Chemistry, Physics, Programming
